Nuclear Compact Reactors

p. 2/ 24

Nuclear Compact Reactors

MONTE-CARLO COUPLED DEPLETION CODES EFFICIENCY FOR RESEARCH REACTOR DESIGN

E. Privas, C. Bouret, S. Nicolas, L. Manifacier

December 5th – 18th IGORR Conference 2017

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.3/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Introduction

TechnicAtome: specializes in the design, construction, operation and maintenance of

compact nuclear reactors

Early stages of core design and industrial studies require a quick and efficient

calculation of key neutronic parameters at any time

Mainly achieved by deterministic calculation schemes

COCONEUT (COre COnception NEUtronic Tool)

Nevertheless

Deterministic codes: problem dependent / V&V process for various kinds of cores

Improvement of CPU power: Monte-Carlo burnup calculations for industrial studies

Aims of this paper:

Monte-Carlo burnup codes for industrial studies (TRIPOLI4®, MCNP, Serpent) ?

Describe a case study part of the V&V process undergone by COCONEUT

Case study:

A multipurpose dummy core designed by TechnicAtome

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.4/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Contents

Depletion calculation methods

Codes used in this study

Case study: Dummy core

Results and Analysis

Conclusion and Outlooks

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.5/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Contents

Depletion calculation methods

Codes used in this study

Case study: Dummy core

Results and Analysis

Conclusion and Outlooks

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.6/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Depletion calculation methods (1/2)

𝝋 𝝋

Step 1 Step 2

𝝋 Step 3

Depletion solver

New compositions

calculation Material

balance 2

Depletion solver

Material

balance 3

1st or 2nd order

methods

Deterministic

General diagram for depletion calculation

Deterministic / stochastic calculation ?

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.7/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Depletion calculation methods (2/2)

Deterministic approach

Fast method for flux calculation industrial studies

Cross sections collapsing

Self shielding

Spatial mesh

Time related mesh

Geometry dependent

Approximations / biases to quantify

Stochastic approach

Exact 3D geometry

Punctual XS for flux calculation

Slower than deterministic calculation

Statistical uncertainties

Spatial mesh for depletion

Time related mesh

Results depending on statistical

convergence

Uncertainties propagation

Time consuming

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.8/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Contents

Depletion calculation methods

Codes used in this study

Case study: Dummy core

Results and Analysis

Conclusion and Outlooks

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.9/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Codes used in this study: MC codes

TRIPOLI4®

MCNP Serpent

• International reference

• Code largely benchmarked

• Many applications at TechnicAtome

• Assessment of JHR neutronic

performances

• Fast

• New methods (perturbation, coupled physic…)

• Automatic mesh

• Undergoing V&V process

• Code developed by CEA (French Alternative

Energies and Atomic Energy Comission)

• Safety studies reference at TechnicAtome

• Polyvalent code

• Large V&V process

• Root based interfaces (pre / post processing)

• Geometry modification during depletion

• Refueling module

• Possibility to develop a tool for

uncertainties propagation

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.10/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Codes used in this study: COCONEUT scheme

1) XS generation Standard FA

Supercritical pattern for Absorber FA

2) Core calculation 2D model (exact)

3D model

Validation: fuel pattern and core

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.11/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Contents

Depletion calculation methods

Codes used in this study

Case study: Dummy core

Results and Analysis

Conclusion and Outlooks

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.12/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

AIMS

Comparing methodologies / non-regression tests

Education and Training object

Validation and qualification of both calculation and

computational techniques

V&V purposes

DESIGN

Only describe two assembly types

Simple to model

Fuel lattice pattern

Add components

Reflector vessel (heavy water…)

Experimental devices

Add ex-core environment

Case study: Dummy core (1/2)

Fuel (MTR type)

Cladding

Aluminum

Boron

Water

Hafnium

Standard Fuel Assembly (SFA) Absorber Fuel Assembly (AFA)

Depending on

the case

study

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.13/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Fuel assembly lattice configuration (2D)

3 configurations

Temperature: 300°K

Total: 62 depleted medium

Full core configuration (2D)

32 assemblies core

16 SFA / 16 AFA

Reflector / coolant : light water

Fuel lattice: 10 cm

Total: 672 depleted medium

Case study: Dummy core (2/2)

SFA

AFA

Dummy Core for V&V and

Education & Training Purposes at

TechnicAtome: In and Ex-Core

Calculations

S. Nicolas, A. Noguès, L. Manifacier,

L. Chabert

For more information

SFA AFA pattern

(rod out) AFA pattern

(rod in)

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.14/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Contents

Depletion calculation methods

Codes used in this study

Case study: Dummy core

Results and Analysis

Conclusion and Outlooks

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.15/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Results and analysis: Standard FA (1/3)

Benchmark considerations

The same consistent parameters are taken into

account for each code and simulation

• Reflecting surface are defined as boundary conditions

• 50 burnup steps with a maximum value of 100 GWd/tU

• Assembly power of: 1.5625 MWth

• Depletion in fuel and boron plates

• Temperature: 300 K

• JEFF-3.1.1 nuclear data library

𝝋𝑬 𝒅𝑬𝑬

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.16/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Results and analysis: Standard FA (2/3)

Initial kinf Monte Carlo codes are all the same within the 1σ range / +150 pcm bias for COCONEUT

Depletion:

Maximum reactivity peak at ~64 GWd/tU (less than 10% of [10B] remains)

COCONEUT: Maximum reactivity discrepancy is found when 10B is half consumed (+280 pcm)

Bias relatively constant between MCNP and Serpent even after 60 GWd/tU

Discrepancies become visible after 60 GWd/tU between TRIPOLI4® and other MC codes

Kinf comparison Reactivity comparison (mean value of the three MC

codes as reference)

Standard FA: multiplication factor comparison

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.17/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Results and analysis: Standard FA (3/3)

[135Xe]: Less than 2.5 % (all codes)

[ 10B]: at 60GWd/tU (10% 10B remains)

MC codes: until 5.0 %

COCONEUT: until 5.5 %

[ 235U]: at the end of depletion

Serpent : 0.5 %

TRIPOLI4®: 1.0 %

COCONEUT: 1.5 %

Good agreement between the codes

Whole concentrations comparison: MCNP as reference

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.18/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Benchmark considerations

The same consistent parameters are taken into

account for each code and simulation

• Reflecting surface on Z axis

• 50 burnup steps with a maximum value of 80 GWd/tU

• Core power of: 50 MWth

• Temperature: 300 K

• JEFF-3.1.1 nuclear data library

Results and analysis: 2D full core (1/3)

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.19/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Results and analysis: 2D full core (2/3) Core calculation: multiplication factor comparison

MC codes

Discrepancy between -110 pcm and +78 pcm

Simulation time: Serpent faster than TRIPOLI4®

COCONEUT vs mean of MC codes Fresh fuel : -235 pcm

Constant bias during the depletion: (between -395 pcm and -235 pcm)

6 factor formula has to be calculated during the depletion to determine

compensations

Slight discrepancy between the

codes.

Next step:

3D core calculation and critical

configurations research during

the depletion.

Kinf comparison Reactivity comparison (mean value of MC codes as

reference)

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.20/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Results and analysis: 2D full core (3/3)

[135Xe] discrepancies

MC codes: less 0.6 %

COCONEUT vs MC: Maximum of 1.6 %

[ 149Sm]

MC codes: less 0.5 %

COCONEUT vs MC: 4.0 % during the first steps

[ 10B]

MC codes: Maximum of 2.5 %

COCONEUT vs MC: Close to 4.5 %

Good agreement between the codes

Whole concentrations comparison

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.21/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Contents

Depletion calculation methods

Codes used in this study

Case study: Dummy core

Results and Analysis

Conclusion and Outlooks

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.22/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Conclusions

Good agreements between MC codes for 2D assembly and 2D full core

Serpent faster than other MC codes

Quantify the differences between normalization methods

Small discrepancy between COCONEUT and MC codes

Constant bias around -300 pcm during the entire depletion for 2D full core

Dummy core is well suitable for core calculation studies and gives a better

understanding of design purpose

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.23/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Outlooks

Optimize MC coupled depletion codes

Adapt time mesh with the flux gradient

Test of refueling algorithm proposed by Serpent and TRIPOLI4®

Changing depletion mesh step by step during the depletion (3D calculation)

Perform uncertainties propagation (compositions / flux)

Comparison with experimental core data

COCONEUT

Estimate compensations with 6 factors formula

281 groups calculation

Perform self shielding during the depletion

Dummy core: future works on new methods for neutron propagation from core to ex-core system

18th IGORR Conference – 3-7 December 2017 Nuclear Compact Reactors

©T

echnic

Ato

me (

avec a

imable

auto

risatio

n d

u C

EA

)

p. 24/ 24

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.25/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

Depletion calculation codes and benchmark

considerations (1/2)

Neutronic analysis :

Rise to equilibrium / Material balance

Flux / power distribution, Absorbers worth

Mainly used for export fuel assembly burnup compositions

to MC codes

Principal model consideration

1) XS calculation

• XS collapsing: MOC calculation 281 26 groups

• Self-shielding performed at the initial step

• AFA is treated as a supecritical pattern representative of neutronic spectrum

in the AFA.

2) Core calculation

• Transport theory (26-group) - 2D exact APOLLO2

• Diffusion theory (4-group) - 3D model CRONOS2

Currently undergoing a large V&V process

Part of this process: estimate the impact of main assumptions on

depleted composition with MC codes

Monte-Carlo Coupled Depletion Codes Efficiency for RR - p.26/ 24 This document is the property of “Société Technique pour l’Énergie Atomique“ and may not

be reproduced or communicated to a third party without prior authorisation

COCONEUT : discrepancy analysis

Results and analysis: Standard FA (3/4)

Relative concentration comparison (%) between COCONEUT

and mean value of MC codes

Main concentrations

Less than 3% discrepancy

Constant bias on 148Nd (burnup indicator)

[10B]

Burned faster with TRIPOLI4® and COCONEUT

depletion chain / power normalization ?

COCONEUT Outlook

Depletion with 281 group

Self shielding during depletion (several steps)